Past Events

Personalized Treatment Plans with Multivariate Outcome Measures - Dr. Somnath Datta

DCEG Events

In this talk, the speaker proposes a novel method for individualized treatment selection when the treatment response is multivariate. This method covers multiple (any number) of treatments and it can be applied for a broad set of models. The key idea to handle multivariate response is to employ a rank aggregation technique to estimate an ordering of treatments based on ranked lists of treatment performance measures such as smooth conditional means and conditional probability of a response for one treatment dominating others. An empirical study demonstrates the performance of the proposed method in finite samples. Finally, the procedure will be illustrated using a real-life dataset on a HIV clinical trial.

NIH Human Research Protections Program Approval – Implications for DCEG Studies - Dr. Sharon Savage

DCEG Events

As of January 1, 2019, there is no longer a requirement for investigators to submit a request for determination by the NIH Office of Human Subjects Research (OHSR) when they are conducting research activities with only de-identified data or biospecimens and have no access to identifiers or the code key. Dr. Savage will present the implications of this new policy for both new and currently ssIRB/OHSRP approved DCEG studies. All are highly encouraged to attend to learn about the new DCEG internal processes.

Computationally-Efficient Framework for Inferring the Evolutionary History of a Tumor Given Single-Cell Sequencing Data - Dr. Julia Chifman

DCEG Events

The advent of single-cell sequencing provides the ability to model clonal evolution of tumors within individual patients. Inference of such within-patient tumor phylogenies has the potential to advance the authors understanding of the variation in the process of tumor progression, with strong clinical implications. Here the authors consider the problem of rapid and robust inference of the tumor phylogeny from a sample of single-cell genotype data under a Markov model that incorporates error in the observed sequencing process.

Malignant Rhabdoid Tumor: A Model for Tumor Immunogenicity Beyond Somatic Mutation Burden - Dr. Josh Waterfall

DCEG Events

Despite having one of the lowest somatic mutation burdens of any type of cancer, we show that rhabdoid tumors are highly infiltrated in many patients. Integrative profiling of human tumors and genetically engineered mouse models of rhabdoid tumor, including single cell RNA and TCR sequencing, highlight immunosuppressive macrophage populations and clonally expanded resident memory T cells.

Use of Bayesian Non-Parametrics for Problems in Causal Inference - Dr. Mike Daniels

DCEG Events

The authors propose a general Bayesian nonparametric (BNP) approach to causal inference in the point treatment setting. The joint distribution of the observed data (outcome, treatment, and confounders) is modeled using an enriched Dirichlet process. The combination of the observed data model and causal assumptions allows us to identify any type of causal effect - differences, ratios, or quantile effects, either marginally or for subpopulations of interest.

2nd NCI-Curie Institute Meeting

DCEG Events

Genetics and Genomics of Uveal Melanoma - Dr. Marc-Henri Stern

DCEG Events

The Future of Low-Dose Radiation Research in the United States

DCEG Events

The symposium will provide an open forum for a national discussion on the value of a long-term strategy to guide low-dose radiation research in the United States and the pros and cons of different options for organizing such research.

Tumor Genomes Shed Light into Somatic Mutational Processes and Cancer Vulnerabilities - Dr. Nuria Lopez-Bigas

DCEG Events

Somatic mutations are the driving force of cancer genome evolution.

Accessible Artificial Intelligence for Cancer Research

DCEG Events

Artificial intelligence (AI) has finally emerged as a useful tool for big data analytics. This is due to decades of research that have provided powerful algorithms and the availability of plentiful computing resources. Despite these advances, AI is not widely accessible due to a steep learning curve and the expense and black box nature of software provided by commercial vendors such as IBM.